Selecting a cost optimum subset of discrete-value dispersion compensation modules (DV-DCMs) subject to maximum module count from an available set of DV-DCMs is a NP-hard problem. We derive a novel dynamic programming algorithm with pseudo-polynomial time bound and show that DV-DCM cost re-scaling can improve the running time. We derive a novel dynamic programming algorithm with pseudo-polynomial time bound (DV-DCMs) subject to maximum module count from an available set of DV-DCMs is a NP-hard problem and show that DV-DCM cost re-scaling can improve the running time.